SHAP-Assisted Resilience Enhancement Against Adversarial Perturbations in Optical and SAR Image Classification
Bibliographic record
Abstract
The increasing reliance on convolutional neural networks (CNNs) for automatic target recognition (ATR) in critical applications necessitates robust defenses against adversarial attacks, which can undermine their reliability. To address this challenge, this letter proposes a novel classification framework that enhances CNN robustness for ATR under adversarial perturbations. Although CNNs are renowned for their high recognition accuracy, their performance can be compromised by subtle adversarial perturbations designed to deceive the classifier. Our methodology is based on extracting specific features from Shapley additive explanations (SHAP) analysis within and outside the detected target area. These features are then used to train a multinomial logistic regression model using the training labels, and the trained regressor performs the classification. The key strength of our framework relies on robustness enhancement against adversarial attacks, particularly designed by the fast gradient sign method (FGSM). We validate our findings through extensive evaluations using two publicly available datasets: the multitype aircraft remote sensing images (MTARSI) dataset, which contains optical images of various aircraft types, and the moving and stationary target acquisition and recognition (MSTAR) dataset, which contains radar images.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".